Fingerprint Clustering for Forensic Data Indexing

Conventional fingerprint classification methods are designed to identify unknown fingerprints from known persons. Such a supervised classification framework, however, can be awkward when the captured fingerprints are not from the enrolled people. This study proposes an unsupervised fingerprint classification mechanism in lieu of the supervised one. Our aim is to partition a collection of unknown fingerprints into clusters, so that each cluster consists of fingerprints from the same finger. This mechanism is helpful for criminal investigators to identify fingerprints from the same people, even when the people is not enrolled in the database. We formulate the task of fingerprint clustering as a problem of minimizing the clustering errors characterized by the Rand index. The index reaches its minimum value only when each cluster consists exclusively of fingerprints from only one finger and the number of clusters equals the number of fingers involved in the fingerprint collection. Experiments conducted using FVC2002 database confirm the feasibility of the proposed method.